IAB Australia has released a new framework to help the advertising and marketing industry use generative AI more safely, consistently and effectively.
The vendor-agnostic LLM Prompting Framework was developed by IAB Australia’s AI Working Group and is designed for marketers, publishers, creative teams and adtech companies using AI across briefs, planning, content development, campaign analysis, audience segmentation and technical and commercial workflows.
The release comes as AI use shifts from individual experimentation to more structured business adoption across marketing and advertising teams. IAB Australia said inconsistent prompting can pose risks to accuracy, privacy, brand safety, and commercial decision-making.
Five editions for different parts of adland
The framework includes five editions: General LLM Prompting, Buy Side, Sell Side, Creative and AdTech.
Each edition includes practical examples that can be adapted for common workflows, including media strategy, audience segmentation, campaign analysis, sponsorship proposals, RFP responses, creative concepting, technical documentation and privacy-safe data architecture.
Jonas Jaanimagi, Tech Lead at IAB Australia, said digital advertising is moving from manual AI experimentation to enterprise-scale orchestration.
“This LLM Prompting Framework and dedicated guides are designed to be the connective tissue between strategic human intent and autonomous machine execution,” Jaanimagi said.
“It is no longer enough to simply write better prompts; businesses must now manage entire agentic workflows within rigorous, industry-standard guardrails for privacy, brand safety, and transparency.
“The IAB Australia AI Working Group is very keen to ensure that as AI tools become more deeply embedded in media planning, inventory discovery and technical infrastructure they remain interoperable and fully compliant with both the requirements of sophisticated marketers and also the ever-evolving regulatory landscape here in Australia.”
From AI curiosity to everyday use
Mark Titley
Mark Titley, Senior Commercial Data Strategist at Paramount and an AI Working Group member, said developing practical AI prompting guidance required cross-functional collaboration.
“The collective expertise from Laurence Demousselle, June Cheung, Adam Goodman, and Simon Thomas, spanning agency strategy, agentic solutions, enterprise platforms, and creative production, was vital in ensuring this guidance is both technically rigorous and commercially practical for Australian publishers, agencies, creatives, adtech vendors, and brands.
“It was a privilege to work with such a diverse group of experts to help our industry move from AI curiosity to confident, everyday use and we look forward to evolving them as AI capabilities and our industry continue to advance.”
Who it is for
The General LLM Prompting Edition covers foundational prompting techniques for all roles, including work with briefs, images, documents and data.
The Buy Side Edition — designed for brand marketers, media agencies and planning teams, with prompts for media mix modelling, audience segmentation, media strategies, channel plans, campaign analysis and stakeholder investment cases.
The Sell Side Edition — aimed at publishers, broadcasters and sales houses, with prompts for sponsorship proposals, advertiser-ready insights, competitor positioning, ad product development and RFP responses.
The Creative Edition — designed for creative directors, copywriters, art directors, designers, social content creators and creative agencies, with a focus on brand voice, visual metaphors, campaign concepts, platform copy and iterative creative development.
The AdTech Edition — tailored to engineers, product managers, data scientists, solutions architects and commercial teams building digital advertising infrastructure, with prompts for privacy-safe data architecture, technical documentation, product specifications, compliance frameworks and go-to-market strategies.
Main image: Jonas Jaanimagi